Update README.md

README.md

@@ -29,11 +29,9 @@ Lightweight CNN + CTC line-level OCR model for North Sámi (`sme`). Trained from
 | Split | CER | WER | Char acc |
 |---|---|---|---|
 | Språkbanken synthetic (val, 30,738 lines) | 2.04% | 7.60% | 97.96% |
-| Benchmark test set (1,048 lines) | 12.24% | 37.35% | 87.72% |
+| Benchmark test set (1,048 lines) | 12.24% | 37.35% | 87.76% |
 
-The 12.24% CER figure is the headline number reported in the accompanying papers. The lower 2.04% CER is the in-distribution validation result during training (best epoch, 97/100). For reference, sequence-level (full-line exact-match) accuracy at the best epoch was 80.76%.
-
-Note: CTC outputs are length-stable (no autoregressive over-generation), so character accuracy is effectively `100 − CER` for this model.
+Character accuracy is reported as `100 − CER`. The 12.24% CER figure is the headline number reported in the accompanying papers. The lower 2.04% CER is the in-distribution validation result during training (best epoch, 97/100). For reference, sequence-level (full-line exact-match) accuracy at the best epoch was 80.76%.
 
 ## Usage